Skip to content

Order blinded paths by reliability criteria #2912

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Mar 7, 2024
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
50 changes: 50 additions & 0 deletions lightning/src/ln/offers_tests.rs
Original file line number Diff line number Diff line change
Expand Up @@ -263,6 +263,56 @@ fn prefers_non_tor_nodes_in_blinded_paths() {
}
}

/// Checks that blinded paths prefer an introduction node that is the most connected.
#[test]
fn prefers_more_connected_nodes_in_blinded_paths() {
let mut accept_forward_cfg = test_default_channel_config();
accept_forward_cfg.accept_forwards_to_priv_channels = true;

let mut features = channelmanager::provided_init_features(&accept_forward_cfg);
features.set_onion_messages_optional();
features.set_route_blinding_optional();

let chanmon_cfgs = create_chanmon_cfgs(6);
let node_cfgs = create_node_cfgs(6, &chanmon_cfgs);

*node_cfgs[1].override_init_features.borrow_mut() = Some(features);

let node_chanmgrs = create_node_chanmgrs(
6, &node_cfgs, &[None, Some(accept_forward_cfg), None, None, None, None]
);
let nodes = create_network(6, &node_cfgs, &node_chanmgrs);

create_unannounced_chan_between_nodes_with_value(&nodes, 0, 1, 10_000_000, 1_000_000_000);
create_unannounced_chan_between_nodes_with_value(&nodes, 2, 3, 10_000_000, 1_000_000_000);
create_announced_chan_between_nodes_with_value(&nodes, 1, 2, 10_000_000, 1_000_000_000);
create_announced_chan_between_nodes_with_value(&nodes, 1, 4, 10_000_000, 1_000_000_000);
create_announced_chan_between_nodes_with_value(&nodes, 1, 5, 10_000_000, 1_000_000_000);
create_announced_chan_between_nodes_with_value(&nodes, 2, 4, 10_000_000, 1_000_000_000);
create_announced_chan_between_nodes_with_value(&nodes, 2, 5, 10_000_000, 1_000_000_000);

// Add extra channels so that more than one of Bob's peers have MIN_PEER_CHANNELS and one has
// more than the others.
create_announced_chan_between_nodes_with_value(&nodes, 0, 4, 10_000_000, 1_000_000_000);
create_announced_chan_between_nodes_with_value(&nodes, 3, 4, 10_000_000, 1_000_000_000);

let (alice, bob, charlie, david) = (&nodes[0], &nodes[1], &nodes[2], &nodes[3]);
let bob_id = bob.node.get_our_node_id();

disconnect_peers(alice, &[charlie, david, &nodes[4], &nodes[5]]);
disconnect_peers(david, &[bob, &nodes[4], &nodes[5]]);

let offer = bob.node
.create_offer_builder("coffee".to_string()).unwrap()
.amount_msats(10_000_000)
.build().unwrap();
assert_ne!(offer.signing_pubkey(), bob_id);
assert!(!offer.paths().is_empty());
for path in offer.paths() {
assert_eq!(path.introduction_node_id, nodes[4].node.get_our_node_id());
}
}

/// Checks that an offer can be paid through blinded paths and that ephemeral pubkeys are used
/// rather than exposing a node's pubkey.
#[test]
Expand Down
10 changes: 7 additions & 3 deletions lightning/src/onion_message/messenger.rs
Original file line number Diff line number Diff line change
Expand Up @@ -364,13 +364,17 @@ where
network_graph
.node(&NodeId::from_pubkey(pubkey))
.filter(|info| info.channels.len() >= MIN_PEER_CHANNELS)
.map(|info| (*pubkey, info.is_tor_only()))
.map(|info| (*pubkey, info.is_tor_only(), info.channels.len()))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think choosing the right introduction points in order to optimize for a specific metric such as reliability is very related to finding optimal autopilot/attachment strategies. Using the node degree as a proxy metric for reliability seems like a solid first step, especially if we're talking about payment reliability.

A few more thoughts though: Further down the road we could also entertain a few other metrics that could make sense, for example optimizing for betweenness centrality or choosing introduction points according to k-Center optimization. (Some of our simulation results in this paper might be of interest here, see for example Fig. 1). Also, we could account for node uptime (regularity of node announcements?) or latency metrics, that might become available as part of the attributable errors proposal?

Doing anything more than just scoring by node degree might of course require us to pre-calculate and store more per-node metrics, so would probably only make sense if we'd ever come around to introduce a node score, possibly as part of #63?

)
.collect::<Vec<_>>();
peer_info.sort_unstable_by(|(_, a_tor_only), (_, b_tor_only)| a_tor_only.cmp(b_tor_only));

// Prefer using non-Tor nodes with the most channels as the introduction node.
peer_info.sort_unstable_by(|(_, a_tor_only, a_channels), (_, b_tor_only, b_channels)| {
a_tor_only.cmp(b_tor_only).then(a_channels.cmp(b_channels).reverse())
});

let paths = peer_info.into_iter()
.map(|(pubkey, _)| vec![pubkey, recipient])
.map(|(pubkey, _, _)| vec![pubkey, recipient])
.map(|node_pks| BlindedPath::new_for_message(&node_pks, &*self.entropy_source, secp_ctx))
.take(MAX_PATHS)
.collect::<Result<Vec<_>, _>>();
Expand Down